https://ceph.com/docs/master/rados/operations/crush-map/#placing-different-pools-on-different-osds
See "rule ssd-primary"
31 окт. 2013 г. 17:29 пользователь "Alexis GÜNST HORN" <alexis.gunsthorn@xxxxxxxxxxxx> написал:
Hello to all,
Here is my ceph osd tree output :
# id weight type name up/down reweight
-1 20 root example
-12 20 drive ssd
-22 20 datacenter ssd-dc1
-104 10 room ssd-dc1-A
-502 10 host A-ceph-osd-1
0 1 osd.0 up 1
...
9 1 osd.9 up 1
-103 10 room ssd-dc1-B
-501 10 host B-ceph-osd-1
10 1 osd.10 up 1
...
19 1 osd.19 up 1
-11 0 drive hdd
-21 0 datacenter hdd-dc1
-102 0 room hdd-dc1-A
-503 0 host A-ceph-osd-2
20 0 osd.20 up 1
...
27 0 osd.27 up 1
-505 0 host A-ceph-osd-3
35 0 osd.35 up 1
...
42 0 osd.42 up 1
-101 0 room hdd-dc1-B
-504 0 host B-ceph-osd-2
28 0 osd.28 up 1
...
34 0 osd.34 up 1
-506 0 host B-ceph-osd-3
43 0 osd.43 up 1
...
50 0 osd.50 up 1
So, in a word,
I've 2 sorts of OSD hosts : full SSD, and full standard HDD.
Hosts are splited in 2 rooms : A and B.
I've created 2 pools :
vm-sdd and vm-hdd
My goal is this one :
for pool vm-sdd, replicate data between room A and B on SSD hosts, so
here between A-ceph-osd-1 and B-ceph-osd-1
for pool vm-hdd, to the same, but only with HDD hosts.
Here is my rules in my crush map :
rule vm-sdd {
ruleset 3
type replicated
min_size 1
max_size 10
step take example
step emit
step take ssd
step chooseleaf firstn 0 type room
step emit
}
rule vm-hdd {
ruleset 4
type replicated
min_size 1
max_size 10
step take example
step emit
step take hdd
step chooseleaf firstn 0 type room
step emit
}
But but... when i do :
ceph osd pool set vm-hddd crush_ruleset 4
it does not work, i get all my PGs stuck.
Do I have missed something ?
Do you have an idea ?
Thanks a lot for your help.
Best Regards - Cordialement
Alexis
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
